In recent years, game units and simulators equipped with on-board virtual image generation apparatuses which make it possible to display images of objects that move through three-dimensional natural virtual space in response to manual control by a game player such as in a; combat have been developed. Such virtual image generation apparatuses are usually equipped with a virtual image generation apparatus main unit that houses a computer unit for executing stored programs, an input device for sending control signals to the computer unit to instruct it to move objects displayed on the screen within the virtual image, a display for displaying the virtual images generated by the computer unit according to the program sequence, and a sound device for generating sounds according to the program sequence.
Examples of game devices with the architecture described above include those with a combat theme in which a player-controlled object (robot, human, or the like) engages in combat with enemy objects with which the player fights over a terrain created in virtual space (hereinafter termed “virtual terrain”). The objects controlled by the player in such game units attack enemies by shooting at them while hiding behind the obstacles and the like which are provided as part of the virtual terrain.
In order to enable three-dimensional display of motion within the virtual space, an image like that perceived when the virtual space is observed from a prescribed visual point is used. This is accomplished using coordinate conversion for perspective projection, whereby a coordinate system for the virtual space is represented in perspective from a prescribed visual point and projected onto a two-dimensional plane lying in front of the visual point. In many cases, the line of sight which extends from the visual point of the virtual image is oriented towards the player's object so that the object controlled by the player is visible to the player. The object controlled by the player is displayed as a subject located nearlly at the center of the display.
However, since the position relationship between the visual point of the virtual image and the player's object is fixed during movement, obstacles sometimes block the line of sight which extends towards the player's object. In such cases, the player's object is obscured by the obstacle, making it difficult for the player to see the movements of the object and making it difficult to control the object. The loss of ability to control the player's object diminishes the excitement of the game, making the game uninteresting.
Such a case will be described making reference to FIG. 6. When a subject R′ is observed by a virtual camera C′, which serves as the visual point for the virtual image in question, position relationships are sometimes such that the line of sight is blocked by an obstacle O, as shown in FIG. 6A. When the line of sight is so obscured, the subject R is displayed obscured by the obstacle O in the generated virtual image, as shown in FIG. 6B. Thus, the player can no longer determine how best to control the subject R′ which is the object of control.
In order to avoid such occurrences, it is possible, for example to:                (1) not display obstacles; or        (2) display the obstacles with wire frames from the beginning.        
However, adopting methods such as (1) produces a new problem in that, while the player's object is visible, obstacles are not visible. Adopting methods such as (2) makes it possible to see obstacles, but since obstacles are displayed with wire frames even when the subject R is not hidden by obstacles O, the look of the game suffers.
In order to solve such problems, the present invention is intended to provide a virtual image generation apparatus that does not employ the aforementioned methods (1) and (2), and that affords a game that does not suffer from impaired look. Another object of the present invention is to provide a virtual image generation apparatus that correctly determines whether a subject can be displayed overlapping a physical object in virtual space, and which performs appropriate transparent processing to make both the subject and physical object visible, and to a method therefor.